This　article　proposes　a　deep　learning　based　approach　for　cyber　attack　detection　in　the　vehicles.　The　proposed　method　is　constructed　based　on　generative　adversarial　network　(GAN)　classification　to　assess　the　message　frames　transferring　between　the　electric　control　unit　(ECU)　and　other　hardware　in　the　vehicle.　To　this　end,　two　networks　called　generator　(G)　and　discriminator　(D)　will　run　an　adversarial　game　to　fool　each　other.　In　such　a　process,　the　most　optimal　structure　is　found　which　distinguish　between　the　model　normal　behavior　and　abnormalities.　Due　to　the　instabilities　existing　in　the　GAN　model,　a　new　optimization　method　based　on　firefly　algorithm　is　proposed　to　create　a　class　of　generators　in　a　feasible　region,　i.e.　the　discriminator　D.　A　three-stage　modification　method　is　also　devised　to　increase　the　algorithm　population　diversity　and　reduce　the　possibility　of　falling　in　local　optima.　The　performance　of　the　model　is　assessed　on　the　experimental　dataset　recorded　from　the　OBD-II　port　of　an　undefined　vehicle.